Effects of electrothermal vortices on insulator-based dielectrophoresis for circulating tumor cell separation

Transcription

1 Electrophoresis 2018, 39, Arian Aghilinejad 1 Mohammad Aghaamoo 2 Xiaolin Chen 1 Jie Xu 3 1 Department of Mechanical Engineering, Washington State University, Vancouver, WA, USA 2 Department of Biomedical Engineering, University of California, Irvine, CA, USA 3 Department of Mechanical and Industrial Engineering, University of Illinois at Chicago, Chicago, IL, USA Received June 26, 2017 Revised September 5, 2017 Accepted September 8, 2017 Research Article Effects of electrothermal vortices on insulator-based dielectrophoresis for circulating tumor cell separation Insulator-based dielectrophoresis (idep) is a powerful technique for separation and manipulation of bioparticles. In recent years, idep designs using arrays of insulating posts have shown promising results toward reaching high-efficiency bioparticle manipulation. Joule heating (JH) and electrothermal (ET) flows have been observed in idep microdevices and significantly affecting their performances. In this research, we utilize mathematical modeling to study, idep technique and the effects of JH and ET flow on device performance and propose a separation scenario for selective trapping of circulating tumor cells (CTCs). A robust numerical model is developed to calculate the distribution of electric and fluid flow fields in the presence of JH and ET flow, and predict the cells trajectory inside the system. Our results indicate that JH not only induces temperature rise in the system, but also may alter the design idep separation scenario by inducing ET vortices that affect the cell s trajectory. To investigate the impact of JH-induced ET flow characteristics and vortex generation on separation efficiency, we introduce a dimensionless force ratio encompassing the effects of electrical field, drag forces, JH, and ET flow. Interestingly, it was found that ET flows can be used to significantly enhance the separation efficiency, even in higher inlet flow rates. Lastly, the effect of post geometry has been discussed. Keywords: Circulating tumor cell / Dielectrophoresis / Electrothermal flow / Joule heating DOI /elps Additional supporting information may be found in the online version of this article at the publisher s web-site 1 Introduction In recent years, dielectrophoresis (DEP) has emerged as a promising technique for isolation and enrichment of target cells from biological samples. One major application is in isolation of circulating tumor cells (CTCs) for cancer diagnosis and prognosis [1]. DEP working principle is based on using a non-uniform electric field to induce motion in polarizable particles or cells. This technique has been brought into biological applications for the separation of blood cells [2], bacteria [3] and CTCs [4]. Recently, insulator based DEP (idep) has gained popularity due to its Correspondence: Dr. Jie Xu, Department of Mechanical and Industrial Engineering, University of Illinois at Chicago, 842 W Taylor St, Chicago, IL 60607, USA Fax: jiexu@uic.edu Abbreviations: CTC, Circulating Tumor Cell; ET, Electrothermal; idep, insulator-based dielectrophoresis; JH, Joule Heating advantages in fabrication and low cost. In contrast to conventional DEP in which the non-uniform electric field is generated by arrays of electrodes inside the microchannel, idep technique utilizes the combination of remote electrodes outside the channel with insulating obstacles within the device to produce the spatial non uniformities in electrical field. This results in the independency of separation efficiency from cells distance to the electrodes [5]. Previously, there have been different types of insulating obstacles studied for idep such as single insulating constrictions [6], serpentine microchannel [7] and arrays of insulating posts [8]. Because of relatively uniform presence of DEP forces along the microchannel, using arrays of microposts have become popular recently. Additionally, there are several studies investigating the effects of different parameters such as the spacing and shapes of the insulating posts [9,10]. Even though idep works with both DC [11] and AC [12] electric field, idep with AC electric field adds frequency as an extra degree of freedom to the design [13]; a unique feature that empowers separation of cells with minor differences in their intrinsic properties. Color Online: See the article online to view Fig. 1 8 in color.

2 870 A. Aghilinejad et al. Electrophoresis 2018, 39, Although idep is reliable and efficient for manipulating particles in lab-on-a-chip devices, Joule heating (JH) is a major concern when dealing with biological applications [14]. This phenomenon, as a result of resistive power dissipation in conductive mediums, may impact device performance due to variation in the medium s temperature-dependent properties, i.e. conductivity, permittivity, viscosity, and density [15]. As a result of the temperature changes due to the JH, local gradients in conductivity and permittivity would be created in the solution, and consequently cause bulk fluid forces and fluid motion, known as electrothermal (ET) flow. Studying and utilizing ET flow in particle manipulation has been previously conducted for DEP-based devices with embedded electrodes inside the microchip [16]. Recently, JH and induced ET flows have been investigated in the context of idep mostly on single constriction microchannels. Hawkins and Kirby modeled JH and the corresponding ET flows in a single constriction idep device, with a DC-offset, and an AC electric field [17]. Their results demonstrated the presence of ET vorticities around the channel constriction, enhancing entrapment by affecting particles deflection and decreasing fluid velocity. Sridharan et al. used both experimentation and numerical simulations to study the effectsofjhonelectroosmoticflowinanidepmicrodevice, i.e. a constriction microchannel [18]. They found that JHinduced ET flow circulations form near channel constrictions and recirculation vortices exist at AC voltages from 200 V to 600 V. This group also experimentally investigated the JH on the electrokinetic particle transport and manipulation and has developed a transient 3D numerical model to study the JH-induced ET flows on microdevice [19, 20]. Kale et al. has investigated the effects of JH and ET streamlines experimentally by introducing the particle trapping number in single constriction microchannel [21]. Prabhakaran et al. has recently observed ET fluid circulations in polymer microfluidic chip and reported the presence of the fluid circulations at 500 V AC and above [22]. Studying idep device with arrays of insulating posts, Gallo-Villanueva et al. observed reduced particle trapping efficiency due to the decrease in electric field gradient because of JH in the posts region [23]. Despite the valuable research conducted on JH and ET flow in microfluidics devices, few has focused directly on applications related to separation of biological samples, specifically CTC separation. Moreover, the literature still lacks a thorough study that provides a detailed design guideline based on JH and ET flow for idep-based technique. To address these issues, here, we define the problem as to study idep with arrays of insulating posts for trapping-based CTC separation. Specifically, we utilize numerical simulation to study extensively how JH and ET flow affect the system performance and what critical design considerations should be taken into account in decision makings. To achieve this goal, we develop a fully coupled electro-thermo-fluid numerical model to investigate the relationship between the device operation parameters and the modelled JH and ET flows. In addition, particle tracing model is coupled with the developed numerical model to further predict the trajectories of cells in the device. Utilizing such a numerical model platform, we study in detail how different operation parameters such as electric field, inlet velocity, medium properties, and post geometries contribute to JH and ET flows. This study is among the first to focus on the effects of JH and ET flows in idep designs for separation of CTCs from blood cells. It is worth noting that although this research focus on idep with arrays of insulating posts, the developed numerical platform is easily extendible to any idep designs for bioparticle separations. 2 Materials and methods 2.1 Governing equations The dielectrophoresis (DEP) arises from the action of the applied electric field on the dipole moment it induces in an object. In the presence of a non-uniform electric field, the electric forces acting on each side of the induced dipole are unequal, causing a net dielectrophoresis force to act on the object. Induced polarization is the underlying phenomenon on which the DEP technique is based. Charges have to be induced to guarantee the continuity of the normal component of total current density in a conductivity gradient, a phenomenon called Maxwell-Wagner interfacial charge relaxation or structural polarization. The dielectrophoretic force is used here to separate the cells based on their differences in dielectric properties. The time average of this force in an inhomogeneous and time-varying electrical field E, is proportional to the cell volume, as shown below: F DEP = 2 ε m r 3 Re (K CM ) E 2 (1) where ε m is the permittivity of the medium, r the radius of the cell, and Re(K CM ) is the real part of the Clausius-Mossotti factor defined as K CM = ε cell ε m (2) ε cell + 2 ε m where ε = ε j is the complex permittivity considering the permittivity and conductivity of the cell and the medium and also angular frequency ( ) of the electrical field. Clausius- Mossotti factor varies between -0.5 to 1 depending on the balance of complex permittivity of the particle and medium. When ε p ε m, it means that the cell is more polarizable than the medium, the DEP force would be positive (pdep) which makes the particles get attracted to the high electric field zones. Conversely, when ε p ε m, the DEP force would be negative (ndep), meaning particles get forced back from high electric field regions. It is to be mentioned that at specific frequency, called crossover frequency, the Clausius-Mossotti factor would be zero, resulting in no dielectrophoretic force. In this study, the calculations were performed, using breast cancer cell and white blood cells as a model, in view of the major importance of these cells in biotechnology. Here, single shell model is used to consider the cell s thin membrane effect on its overall dielectric performance [24]. In this

3 Electrophoresis 2018, 39, General 871 Table 1. Mechanical and electrical properties of selected cells Cell type Tissue r ( m) mem mem ( S m ) cyt cyt ( S m ) d(nm) Ref. WBC Blood [26, 38] MDA-231 Breast [25, 39] model, equivalent cell complex permittivity can be calculated using cytoplasm and membrane complex permittivity (ε cyt and (ε mem ): ( 3 εcyt ε + 2 mem ε cell = ε mem 3 r cell r cell d ) ε cyt+2ε mem ( εcyt εmem ε cyt+2ε mem ) (3) where = and d is the cell membrane thickness. The dielectric properties of breast cancer cells and granulocytes are reported by Becker et al. [25] and Yang et al. [26], respectively, and are listed in Table 1. Based on these dielectric properties, the real part of Clausius- Mossotti is plotted with respect to different applied electric field frequency and medium conductivities (Fig. 1). According to Fig. 1A, by increasing the frequency, the critical conductivity at which the cells change behavior from pdep to ndep will be increased. As shown in Fig. 1B, MDA-231 and WBC behave differently at different frequencies for two representative conductivities. For the purpose of CTC separation, we take the advantage of such a difference to trap one type of cells while allowing the other ones to pass through the device. To achieve this, if the applied frequency is more than the crossover frequency for MDA-231, pdep will be applied to CTCs. On the other hand, when the frequency is less than the crossover frequency for white blood cell, it experiences ndep; hence in this middle range, breast cancer cells would experience pdep which attracts them to the microposts and possibly results in trapping while white blood cells would experience ndep. Generally, it is easier to attract particles using pdep rather than repelling them by ndep [14]. Furthermore, working in pdep regime is more straight-forward especially in low conductive mediums [27]. Most of the previous studies approximated medium conductivity dependency to temperature as a linear function [15]. However, Porras et al. demonstrate that it is more accurate to use quadratic equation as [28] = 0 [ 1 + P1 (T mean T 0 ) + P 2 (T mean T 0 ) 2] (4) where 0 is conductivity at reference temperature (T 0 )and parameters P 1 and P 2 were determined experimentally by Porras et al. which here would be and 4.3 K respectively. K 2 For the permittivity and viscosity of the medium, same as previous studies, we have [15, 18, 20, 29] ε (T) = ε 0 (1 + (T T 0 )) (5) ( ) 1713K = exp (6) T where ε 0 is fluid permittivity at reference temperature and is temperature coefficient of fluid permittivity which is in this study. Other properties such as thermal conductivity and heat capacity can be safely assumed as constant K as DI water in our study. 2.2 Numerical model description Joule heating modelling with associated ET flows requires coupling three physics: thermal, electric, and fluid flow. In this regard, COMSOL Multiphysics (Fluid Flow, Heat transfer in fluids, and AC/DC) was used to solve for the flow, temperature, and electric fields in the system. Also, we have used particle tracing module in COMSOL Multiphysics to track the trajectory of the bioparticles. The geometry is shown in Fig. 2. For calculating flow rates, we assumed the height of the channel as 100 m. Figure 1. (A) Clausius- Mossoti factor for MDA-231 in different conductivities, (B) The behavior of Breast cancer cell (blue) and White blood cell (red) for different frequencies.

4 872 A. Aghilinejad et al. Electrophoresis 2018, 39, Effect of joule heating on temperature, conductivity, and electrical field Figure 2. (A) Isometric view of the idep device, (B) Geometry of the microchip and design parameters. The equations and numerical method to solve them are illustrated in detail in the supporting document. To solve these coupled physics, three sets of boundary conditions are required, and summarized in Table 2. One of the major phenomena of idep technique is the resulting high temperature gradients that would rise due to the high electric field gradients. In Fig. 4A, the temperature profile on the centerline of the microchannel is depicted. Due to heat flux boundary condition that is applied on the outlet, the temperature first started to increase to its maximum and after that decrease. Figure 4B depicts the maximum temperature inside the channel for different initial conductivities with respect to the applied voltage. Because our goal here was to separate MDA-231 from WBCs, for different conductivities, the frequency should be changed in order to have different type of DEP forces (pdep and ndep). We used 20 khz for 0.01 S/m, 100 khz for 0.05 S/m and 300 khz for S/m and 0.1 S/m conductivities according to Fig. 1. These frequencies are in line with the ranges which were previously 3 Results and discussion 3.1 Single constriction idep; Validation study To validate our developed numerical model, the ET flows, as a result of the non-uniform temperature rise in the channel, were modeled and compared with experimental and numerical results reported by Sridharan et al. [18] (Fig. 3) in the special case of 600 V AC applied voltage. According to the results, our model can predict well the ET flows associated with Joule heating. Specifically, as shown in Fig. 3B, the model can accurately predict the ET vortices with specific circulation directions at inlet and outlet of the microchannel. In order to validate of our simulation, we mainly focus on ET flow pattern and the presence of vortices rather than focusing on magnitude. Sridharen et al. [18] used very high convective heat transfer coefficient in their research in order to be consistent with their experimental results. Therefore comparing the magnitudes would not be helpful to compare our simulation with their results. However, for the current study, the ability of our developed numerical model for predicting both Joule heating and the pattern of ET flows would be sufficient enough. Figure 3. Comparison of the ET flows, as a result of Joule heating, between (A) the work by Sridharan et al., and (B) our developed numerical model. The results shows the capability of our model to predict such phenomenon. Table 2. Summary of the utilized boundary conditions for modeling Joule heating and the corresponding electrothermal flow Physics Inlet Outlet Channel walls Electric field Electric potential (V = V in ) Ground (V = 0) Electric insulation (n ( E) = 0) Fluid flow Velocity inlet (v = U in ) Pressure outlet (P out = 0) No slip (v = 0) Heat transfer (T = T 0 ) Outflow ( n (k T ) = 0) Constant Temperature T = T 0 (Room Temperature)

5 Electrophoresis 2018, 39, General 873 Figure 4. (A) Temperature profile on the centerline of the microdevice, (B) Maximum temperature value for different conductivities for common range of applied voltage in idep devices, (C) changes in conductivity due to the temperature gradient inside the microchip, (D) Deviation from the electric field by comparing the gradient of electric field squared when Joule heating is considered ( E 2 JH) neglected ( E 2 ). reported from the experiments done by Das et al. [35]. Note that although we focused on MDA-231 for separation in this study, most of CTCs behave with a similar trend in the presence of electrical field. Therefore the strategy deployed in this research, can also be applied for different types of CTCs. In Fig. 4C, the maximum amount of conductivity for different applied voltages has been presented in different initial conductivities. Even for moderate voltage differences, there would be a significant change in conductivity. Such a highly non-linear trend and its effects should be considered in the device behavior specially for trapping the particles as it alters the Classius-Mossoti factor. Figure 4D represents the variation in E 2 when JH is considered ( E 2 JH)bycomparing the values obtained by neglecting JH ( E 2 ). By considering these changes in conductivity and electric field gradient due to the temperature in design step, we would be able to predict the influence of the JH on electrical performance of the device. 3.3 Effect of fluid circulations on particle separation Presence of ET forces inside the device causes motion in the fluid and creates vortices. By modeling the ET force in the microchip, we can predict these motions inside the microchannel. An important implication of these modeling would be to enable the designers to control the effects of ET flow and even use these vortices inside their devices for separation. Figure 5 shows an example case illustrating the JH-induce ET flows in the microchip with an applied voltage of 500 V. The presence of these vortices could change the regular trajectory of the particles that is predicted by DEP force. For quantitatively showing the effect of these vortices on the main flow, the velocity profile is depicted for two different cases on the centerline: one that considers ET flows and one that neglects ET flows (Fig. 5B and 5C). Note that the velocity profile is depicted for two cases as well, one with 0.05 S/m conductivity for media and another with 0.01 S/m conductivity. It is clear that in lower conductivity, the effect of vortices is negligible (Fig. 5C), as previously reported by Gallo-Villanueva et al. [23] while the effect of ET flow on higher conductivity (Fig. 5B) is significant. Hereafter, all the further studies are conducted in a medium with 0.05 S/m conductivity because of investigating the effects of ET vortices in physiological range on particle separation. To investigate the effect of ET flow on particle trajectory, the trajectory of particles was modeled in the absence and presence of ET flow (Fig. 6). The inlet velocity of the flow was 500 m/s which is among the highest inlet velocities for idep devices corresponded to 0.1 ml/h flow rate. Figure 6 represents the difference in particle trapping in the presence of electrothermal flow. Comparing Fig. 6A and 6B, presence of ET flow decreases the velocity of particles in the first columns of posts and cause more particles to trap.

6 874 A. Aghilinejad et al. Electrophoresis 2018, 39, Figure 5. (A) Fluid streamlines inside the microchannel which causes vortices to arise due to electrothermal forces between the posts in a medium with 0.05 S/m conductivity, velocity profile on the centerline (B) For 0.05 S/m initial conductivity, (C) For 0.01 S/m initial conductivity. caused by the fluid flow due to the ET and pressure gradient effects. This force is calculated based on Eq. S-10 considering only the fluid velocity and neglecting the cell velocity. To trap the target cells, higher value of the force ratio experienced by those cells is more desirable. The main advantage of using this parameter is to simultaneously investigate the effects of temperature and JH on electrical field, conductivity, ET force and velocity which could define the functionality and performance of the design. Using Eqs. (1) and S-10,, the force ratio, would be Figure 6. Simulation results obtained using COMSOL Multiphysics software. A comparison of cell (MDA-231) trajectory inside the channel: (A) In the absence of ET flow, (B) In the presence of ET flow. Note that more cells are getting trapped in the presence of ET flow. These simulations could show the significance of the effect of ET flow even in high flow rates. As it could be noticed from Fig. 6, in the absence of ET flow, more cells are running away from pdep traps. Hence, the presence of ET flow could enhance the separation of the cells. With respect to this point that significant temperature rise will threaten the cells viability, decreasing the cells residual time in DEP traps compensates this issue while allowing us to take advantage of JH-induced ET vortices. One way to achieve this is to periodically apply perpendicular flow to release the trapped cells [36]. To gain an accurate and quantified comprehension on how JH and corresponding ET flow would affect cell separation and manipulation, we define a dimensionless force ratio as the ratio between the DEP force and surrounding fluid force. The surrounding fluid force used in the force ratio is = F DEP F flow = r 2 cell 3 ( ) εm Re(K CM) E 2 v Figure 7 represents the force ratio along the centerline of the microchip for 0.02 ml ml, 0.04 and 0.06 ml inlet hr hr hr flow rates at 500 V applied voltage in 100 khz to trap the MDA-231 cancer cell. In Fig. 7A to 7C, the effect of ET flow was neglected while in the rests, the ET flow effect was considered. As it is shown in Fig. 7, in all the cases we can achieve higher force ratios by using the ET flow on the centerline of the microchannel at first columns of posts which generally are more important than others for trapping the particles. Due to the presence of ET flow, the magnitude of particles velocity would decrease and cause higher force ratio. Note that force ratio decreases by moving along the centerline toward the outlet. The main reason is the decrease in E 2 due to the JH-induced increase in conductivity according to Fig. 4. Although the force ratio decreases by moving along the channel, we are still capable to get higher force ratios and trap the cells in first rows. (7)

7 Electrophoresis 2018, 39, General 875 Figure 7. Force ratio for (A) 0.02 ml ml, (B) 0.04 and (C) 0.06 ml inlet flow rates and 500 V applied voltage for breast cancer cell. These hr hr hr results show the effect of ET flow which enhances the cell separation by decreasing the cell s velocity. 3.4 Effect of geometry on fluid circulations Post shape and configuration has a direct effect on idep trapping efficiency as it affects electric and flow field distribution throughout the system. In Fig. 8, the comparison for three basic geometries (circle, diamond, square) has been done. As shown in Fig. 8, E 2 for diamond shape post is higher than others as confirmed by previous researches [10, 37]. However, the force ratio for diamond shape microposts is smaller than circular and square shape posts. The reason is that in the presence of the ET vortices inside the channel, the velocity profile is affected by these vortices and shape of microposts has influence on this profile. In Fig. 8D, the maximum force ratio for square shape post is higher than circular and diamond for different inlet flow rates. Note that these values are calculated along the centerline of the device which are located at the largest distance from the posts. This means that trapping of the particles can also happen in values less than one. Figure 8E compares the temperature rise in square and diamond shape post relative to circular one. The negative percent for diamond represents that in these designs, the temperature rise is less than the circular posts. These graphs provide a thorough view for using the different shape post depends to the application and can be served as a design guideline. For instance, dealing with particles which are not temperature sensitive, choosing a square shape Figure 8. (A, B, C) The gradient of the electric field squared for different geometries along the centerline, (D) Force ratio for different geometries in different flow rates, (E) Temperature rise inside the channel for square and diamond shape microposts relative to circular post.

8 876 A. Aghilinejad et al. Electrophoresis 2018, 39, post would provide higher efficiency and trapping ratio while dealing with temperature sensitive bioparticles such as mammalian cells, diamond would be the best option for keeping the temperature low and does not decrease the force ratio significantly compare to circular posts. 4 Concluding remarks In this study, mathematical modeling was used to simulate the idep device performance and design scenario for CTC trapping and separation. To achieve a comprehensive conception toward the technique, JH and induced ET flows was modeled and studied. We developed a fully coupled electrothermo-fluid numerical tool via temperature dependent material properties and particle tracing modeling to predict the cell trajectory and device performance for different conditions. We also demonstrated that by choosing appropriate frequency, the selective trapping of CTCs become possible, according to its dielectric properties. The implication of our finding is that ET flows are not always a disadvantage but also can be used to enhance the performance for separation. To quantify this effect, dimensionless force ratio was introduced and utilized in the modeling to show the effects of ET flow and facilitate the comparison between different embedded insulating post geometries which is very critical for idep devices. Results show that although the diamond shape posts provide higher electric field gradient, square shape microposts have higher force ratio in presence of ET flows. The systematic research work conducted here provides the fundamental groundwork as well as design guidelines for developing efficient idep devices for bioparticles and cells separation. The authors have declared no conflict of interest. 5 References [1] Shields, C. W., Reyes, C. D., Lopez, G. P., Lab on a Chip 2015, 15, [2] Cheng, I. F., Froude, V. E., Zhu, Y. X., Chang, H. C., Chang, H. C., Lab on a Chip 2009, 9, [3] Lapizco-Encinas, B. H., Davalos, R. V., Simmons, B. A., Cummings, E. B., Fintschenko, Y., J. Microbiological Methods 2005, 62, [4] Moon, H. S., Kwon, K., Kim, S. I., Han, H., Sohn, J., Lee, S., Jung, H. I., Lab on a Chip 2011, 11, [5] Srivastava, S. K., Gencoglu, A., Minerick, A. R., Analyt. Bioanalyt. Chem. 2011, 399, [6] Barbulovic-Nad, I., Xuan, X. C., Lee, J. S. H., Li, D. Q., Lab on a Chip 2006, 6, [7] Zhu, J. J., Tzeng, T. R. J., Hu, G. Q., Xuan, X. C., Microfluid. Nanofluid. 2009, 7, [8] Shafiee, H., Sano, M. B., Henslee, E. A., Caldwell, J. L., Davalos, R. V., Lab on a Chip 2010, 10, [9] Mohammadi, M., Zare, M. J., Madadi, H., Sellares, J., Casals-Terre, J., Analyt. Bioanalyt. Chem. 2016, 408, [10] LaLonde, A., Romero-Creel, M. F., Lapizco-Encinas, B. H., Electrophoresis 2015, 36, [11] Lapizco-Encinas, B. H., Simmons, B. A., Cummings, E. B., Fintschenko, Y., Analyt. Chem. 2004, 76, [12] Nakidde, D., Zellner, P., Alemi, M. M., Shake, T., Hosseini, Y., Riquelme, M. V., Pruden, A., Agah, M., Biomicrofluidics 2015, 9, [13] Cetin, B., Li, D. Q., Electrophoresis 2011, 32, [14] Voldman, J., Ann. Rev. Biomed. Engin., Annual Reviews, Palo Alto 2006, [15] Xuan, X. C., Electrophoresis 2008, 29, [16] Sin, M. L. Y., Shimabukuro, Y., Wong, P. K., Nanotechnology 2009, 20, 9. [17] Hawkins, B. G., Kirby, B. J., Electrophoresis 2010, 31, [18] Sridharan, S., Zhu, J. J., Hu, G. Q., Xuan, X. C., Electrophoresis 2011, 32, [19] Zhu, J. J., Sridharan, S., Hu, G. Q., Xuan, X. C., J. Micromechan. Microengin. 2012, 22, 6. [20] Kale, A., Patel, S., Hu, G. Q., Xuan, X. C., Electrophoresis 2013, 34, [21] Kale, A., Patel, S., Qian, S. Z., Hu, G. Q., Xuan, X. C., Electrophoresis 2014, 35, [22] Prabhakaran, R. A., Zhou, Y. L., Patel, S., Kale, A., Song, Y. X., Hu, G. Q., Xuan, X. C., Electrophoresis 2017, 38, [23] Gallo-Villanueva, R. C., Sano, M. B., Lapizco-Encinas, B. H., Davalos, R. V., Electrophoresis 2014, 35, [24] Morgan, H., Sun, T., Holmes, D., Gawad, S., Green, N. G., J. Phys. D-Appl. Phys. 2007, 40, [25] Becker, F. F., Wang, X. B., Huang, Y., Pethig, R., Vykoukal, J., Gascoyne, P. R. C., Proc. Natl. Acad. Sci. U. S. A. 1995, 92, [26] Yang, J., Huang, Y., Wang, X. J., Wang, X. B., Becker, F. F., Gascoyne, P. R. C., Biophys. J. 1999, 76, [27] Aldaeus, F., Lin, Y., Roeraade, J., Amberg, G., Electrophoresis 2005, 26, [28] Porras, S. P., Marziali, E., Gas, B., Kenndler, E., Electrophoresis 2003, 24, [29] Knox, J. H., McCormack, K. A., Chromatographia 1994, 38, [30] Stubbe, M., Gimsa, J., Coll. Surfaces a-physicochem. Engin. Aspects 2011, 376, [31] Stubbe, M., Gyurova, A., Gimsa, J., Electrophoresis 2013, 34, [32] Ramos, A., Morgan, H., Green, N. G., Castellanos, A., J. Phys. D-Appl. Phys. 1998, 31, [33] Li, Y. B., Ren, Y. K., Liu, W. Y., Chen, X. M., Tao, Y., Jiang, H. Y., Electrophoresis 2017, 38, [34] Liu, W. Y., Ren, Y. K., Shao, J. Y., Jiang, H. Y., Ding, Y. C., J. Phys. D-Appl. Phys. 2014, 47, [35] Das, C. M., Becker, F., Vernon, S., Noshari, J., Joyce, C., Gascoyne, P. R. C., Analyt. Chem. 2005, 77,

9 Electrophoresis 2018, 39, General 877 [36] Prieto, J. L., Lu, J., Nourse, J. L., Flanagan, L. A., Lee, A. P., Lab on a Chip 2012, 12, [37] LaLonde, A., Gencoglu, A., Romero-Creel, M. F., Koppula, K. S., Lapizco-Encinas, B. H., J. Chromatogr. A 2014, 1344, [38] Gascoyne, P. R. C., Shim, S., Noshari, J., Becker, F. F., Stemke-Hale, K., Electrophoresis 2013, 34, [39] Gascoyne, P. R. C., Wang, X. B., Huang, Y., Becker, F. F., IEEE Trans. Ind. Appl. 1997, 33,